Volume 12 Issue 1 ( March 2025)

According to U.S. Energy Information Administration (EIA) energy consumption and carbon emission are projected to increase globally. In metal manufacturing there is continuous increase in energy demand and carbon emission. Machining is one of them which is widely used in discrete manufacturing sector. The performance of the machining process depends upon cutting parameters. These parameters are selected based on various criteria like minimum cost, maximum production rate, maximum material removal rate, and minimum energy consumption. For this, experiments have been conducted and based on collected data of input parameters cutting speed ,feed and depth of cut and corresponding energy consumption, modeling has been done. In the present work algorithms related to machine learning have been applied to the experimental data for modeling to predict energy consumption The performance of algorithms has been compared using mean squared error (MSE), mean absolute error (MAE) and R2 value. The four favorable algorithms have been compared in terms of accuracy with the experimental data. Among the models tested, the Lasso Regression with Cross-Validation performed the best with lowest mean squared error (MSE) 158.9, mean absolute error (MAE) 8.37 and highest R2 value 0.991.

Keywords: Machine learning; Data set; Regression models; Minimum energy criteria; Machining process; Energy consumption; Performance evaluation